“The medical and social problems of Indigenous people are certainly real, but should residential schools be indicted as the chief or even a major cause? The research is poor and the evidence unconvincing.”
By Tom Flanagan
“TRAUMA” IS A GREEK WORD meaning wound or injury. Its original and still current use in medicine is to describe physical injuries, such as whiplash after a car accident. More recently, it has been widely used to describe psychological stress caused by wars and natural catastrophes, i.e. Post-Traumatic Stress Syndrome. An even further and more recent extension has applied the concept of trauma to whole peoples or nations, as in the phrase “historical trauma.”
Historical trauma is not a bad way to describe what happened to the original peoples of what is now Canada. The coming of the Europeans brought killer diseases against which they had no inherited immunity. It eroded their traditional way of life and made them a small demographic minority in a new nation. Of course, this did not happen all at once. The fur trade was a boon to the Indians and Inuit in much of Canada. By bringing new technologies and opportunities for trade, it raised their standard of living, even though possession of new weapons also touched off new wars. But in the long run the advance of European civilization was devastating for the natives, as agriculture, resource exploitation, and industrial production became the motors of the new economy. Indians found themselves pushed into Indian reserves, where they were protected from extermination but were also made remote from economic progress.
So there is no question that history dealt the First Nations a bad hand. But it is another question whether residential schools are particularly responsible for their plight. It is commonly asserted today that residential schools are responsible for the “intergenerational trauma” of Indigenous people by having cut them off from their traditional languages and cultures. It is further said that the schools, by removing children from their parents, damaged their sense of family life, so that they were unable to become good parents when it was their turn to marry and raise families. Thus the damage was passed from generation to generation, resulting in the ills that plague Indigenous people today: lower life expectancy, higher disease morbidity, alcohol and drug abuse, unemployment, crime, and so on.
The legacy media routinely present the thesis of intergenerational trauma from the residential school experience as if it were an established fact. A typical example is a CBC story from 2021 entitled “How residential school trauma of previous generations continues to tear through Indigenous families: Trauma can have physical and mental effects for six generations, Indigenous health experts say.” The format of such stories has become stylized: a brief, highly torqued portrait of residential schools, a few quotes from “experts,” and a couple of personal accounts. But at least one of the stories is far from heart-rending. It concerns a woman who is now a tenured professor of public health at the University of Toronto. She says her parents went to residential school, and she was largely raised in foster homes. "My kids have to deal with this sort of, sometimes crazy, sometimes unstable mom, because I didn't have parents because they went to residential school," she says, adding that she didn’t like to go to parent-teacher conferences for her four children because they “evoked negative feelings.”
But by any reasonable standard, this woman’s life ought to be considered an inspiration, not a disaster. After a difficult childhood, she earned a Ph.D. in child psychology, now holds a tenured position at one of Canada’s most prestigious universities, and seems to have successfully (she doesn’t say otherwise) mothered four children. She would be a viable candidate for super-woman! Whatever intergenerational traumas she claims to have suffered, they haven’t kept her from achieving extraordinary success.
Behind these superficial stories in the media lies a body of academic research, the findings of which need to be scrutinized more carefully. There are numerous methodological reasons to be cautious about what is reported in this literature.
First, with a tip of the deerstalker hat to Sherlock Holmes, there is a dog that didn’t bark. The obvious first step in research should have been to compare the life outcomes of those who attended residential school to those who did not. Were there differences in variables such as income, employment history, educational attainment, quality of housing, family stability, use of drugs and alcohol, obesity, and other objectively diagnosable medical conditions? Among those who attended residential school, a further investigation, on the dose-response model in medicine, would be to see if a longer period of attendance was correlated with more negative outcomes, as would be predicted by the standard indictment of the schools.
Unfortunately, almost none of this research has been done. A golden opportunity was missed in the First Nations Regional Health Survey (RHS), carried out in the years 2008-2010. In this major effort funded by a consortium of government agencies, researchers carried out more than 21,000 interviews on Indian reserves across Canada. Reserves and individuals were randomly selected in a two-stage sampling framework. The survey included questions about residential school attendance as well as the objective life-outcome variables mentioned above. Yet the researchers published few comparisons of objective outcomes of IRS attendees against others, even though the report is critical of residential schools in several places. They did say that those who attended residential school were more likely to report having been diagnosed with at least one chronic health condition, but that finding means little because attendees were much older on average than other respondents, and age is notoriously associated with chronic health problems. The report contains nothing on how IRS attendees performed with respect to income, employment, housing, and other obviously important, objectively measurable variables.
THIS NEGLECT OF THE obvious has been typical of the Canadian research of the alleged IRS factor in intergenerational trauma. A comprehensive literature review published in 2017 lists 67 published studies on the subject. This research, carried out almost entirely by social workers, psychologists, psychiatrists, physicians, and nurses, reports virtually nothing about the income, employment, education, and other objectively measurable characteristics of IRS attendees. There has been little work by economists, sociologists, and political scientists, who might have been more interested in objective outcomes. That dog simply didn’t bark.
One exception is the work of Simon Fraser University economist Donna Feir, who found that Indian mothers who attended IRS had a similar socio-economic status to those who did not. More recently, she and a colleague showed in a highly sophisticated analysis that, at least from the 1950s onward, those who attended residential schools tended to experience an increase in height, decrease in obesity, and lesser prevalence of diabetes, in comparison to those who went to non-residential schools. Substantial statistical acrobatics were required to tease out these results because children sent to residential school in this period tended to be shorter, fatter, and in poorer health than those who went elsewhere because the schools served in effect as refuges for abused and neglected children.
Apart from Feir’s work, published research deals almost exclusively with subjective variables such as self-reported health and happiness among residential school attendees and their children (and grandchildren). The latter is an increasingly important part of the research. Almost all the schools were shut down by the 1980s, so most who attended are deceased or at least aged. Researchers, therefore, are focussing on their children or even grandchildren, on the assumption that negative effects of IRS attendance might be passed down across generations. Generally speaking, respondents to questionnaires and interviews are sorted as to whether they had at least one parent, or perhaps one grandparent, who attended IRS.
It is an interesting approach, but it neglects other factors of family life that are politically incorrect to mention today but that are obviously important from a common-sense viewpoint. For example, did the child come from a stable, two-parent family? Was there a mixed marriage in which a non-Indian parent and relatives could provide additional opportunities for children?
Here is how the authors of the review article summarize some of the results of the 67 articles they have collected:
Twelve papers used self-reported health or general quality of life as an outcome measure and found that people who had attended residential schools generally felt as though their health or quality of life had been negatively impacted…. However, while the studies reveal negative effects in relation to the residential school system, this cannot be said for everyone who attended. For example, some studies have found better overall reported health among those with family members who attended. … Physical health problems, namely chronic health conditions and infectious diseases, were also apparent in the literature. Thirteen papers related specific physical health conditions to residential school attendance. These included conditions such as HIV/AIDS, chronic conditions (e.g. diabetes, obesity), tuberculosis (TB), Hepatitis C virus (HCV), chronic headaches, arthritis, allergies, and sexually transmitted infections (STIs). … [M]any First Nations people who had personally attended residential schools reported suffering from physical ailments including, chronic headaches, heart problems, and arthritis.
Mental health, and particularly emotional well-being, was the area of health most commonly identified as affected by residential school attendance. Forty-three studies reviewed found that personal or intergenerational residential school attendance was related to mental health issues such as mental distress, depression, addictive behaviours and substance misuse, stress, and suicidal behaviours. … Familial residential school attendance has been associated with lower self-perceived mental health and a higher risk of distress and suicidal behaviours. … Intergenerational effects were found [by one researcher] among women who had parents or grandparents attend residential schools, with women reporting that familial attendance at residential school had had an enduring impact on their lives and mental health.
There are some obvious problems with these findings that can be stated by anyone familiar with statistical research in the social sciences. An inventory of such problems can be found in the famous article by Stanford medical researcher John Ioannidis, “Why Most Published Research Findings are False,” so I draw on Ioannidis for the following critique.
First, these studies were all carried out in the first two decades of the 21st century, after the Report of the Royal Commission on Aboriginal Peoples (1996) condemned Indian Residential Schools. After that, the government of Canada negotiated the Indian Residential Schools Settlement Agreement (2006), leading to compensation payments of $5 billion to $6 billion to those who had attended the schools. Then the Truth and Reconciliation Commission held widely publicized hearings from 2010 to 2015, at which thousands of people who had attended residential schools were encouraged to testify about their experiences, without much balancing testimony solicited from those who had taught or otherwise worked in the schools. So for more than two decades, former attendees heard over and over that the schools were the worst thing that ever happened to them, priming them to tell researchers that residential schools were a major contributor to whatever problems they experienced in their own lives. And a great deal of money was at stake. As Ioannidis puts it, “The greater the financial and other interests and prejudices in a scientific field, the less likely the research findings are to be true.”
Another problem is the design of these studies. When multiple researchers start with a single factor — in this case, attendance at Indian residential schools — and then look for association with a range of variables, some positive correlations are bound to show up. The problem is magnified when dealing with self-reported social and psychological variables such as stress, anxiety, suicidal thoughts, general good health, drug and alcohol abuse, depression, self-esteem, fear, resentment, shame, and troubled relationships, which are often defined by different researchers in varying ways. The lack of rigorous, universally accepted definitions makes it easier to squeeze positive results out of ambiguous data. To quote Ioannidis again, “The greater the flexibility in designs, definitions, outcomes, and analytical modes in a scientific field, the less likely the research findings are to be true.”
The problem is even further magnified when researchers divide their samples into small sub-samples, as was sometimes done in these studies, proclaiming that a result holds true for only one sex, or only for certain age groups. The finding may be interesting, but the methodology makes it tentative at best. Sub-samples in quantitative research seldom exactly mirror the larger sample, and their smaller size gives rise to larger error variance.
The underlying problem can be illustrated by a simple thought experiment. Picture an unweighted coin. The chances of flipping five heads in a row are 0.5 to the fifth power, or approximately 0.03. That is better than the level of statistical significance of 0.05 used in much research. Yet if you flip the coin a hundred times, you will not be surprised to have a run of five (or even more) heads somewhere along the line. The improbable almost always happens if there are a lot of trials. Similarly, if you run correlations of one variable against many others, you will get some correlations that are really chance phenomena, even if the computer says the finding is statistically significant because the probability of it happening by chance is p <.05. This problem plagues all statistical research and can be overcome only by replication of studies using different samples.
Further false results arise when researchers are seeking positive correlations and expect to find them, which is overwhelmingly true of the literature on the effects of attendance at residential schools. All the publications start by giving a negative portrait of the schools, which leaves no doubt where the sympathies of the researchers lie. This is not to say that they are dishonest in the way they carry out their work, but they are human beings whose sympathies can influence the way they interpret and report ambiguous findings.
Another issue is that a number of these studies are based on small samples that do not allow comparative findings. As Ioannidis puts it, “The smaller the studies conducted in a scientific field, the less likely the research findings are to be true.” Fourteen of the 67 studies summarized in the review article mentioned above were based on sample sizes of fewer than 100, and several others were not much larger. Researchers in most of these small studies interviewed only residential school attendees without a control group. If you talk only to attendees, you may find that many of them mention various social and psychological problems, but that doesn’t show that such problems are more prevalent among them than among those who attended reserve day schools, or public schools in town, or no school at all.
The medical and social problems of Indigenous people are certainly real, but should residential schools be indicted as the chief or even a major cause? Beyond the methodological issues highlighted above, several factual obstacles to accepting the thesis are apparent. Careful tabulations by independent researcher Nina Green show that at most one-third of Indian children ever attended IRS, and in most years the percentage was considerably less. There were always more students in day schools on Indian reserves, plus some who attended residential schools daily while continuing to live at home, or who went to public or Catholic schools in nearby towns, or did not go to any school at all. And of those who did go to residential schools, many were there only briefly; the average period of attendance has been estimated at 4.5 years. Was that really enough to destroy children’s bonds with their family and traditional culture? Other influences upon First Nations seem more important, such as the confinement to Indian reserves, which impeded economic progress, followed by extension of the welfare state, which undermined previously strong Indigenous families.
In any case, the past is over and cannot be changed. Canada has profusely apologised, especially for residential schools but for other forms of education, too. Cash compensation totalling more than $30 billion has been paid out or promised to First Nations individuals for the alleged sins of the past, and more will probably be forthcoming. Indigenous people can now do more for themselves and their children by focussing on things they can change, such as low educational achievement, family disintegration, and governance on reserves, rather than rehearsing the past.
 Jared Diamond, Guns, Germs, and Steel (New York: W.W. Norton, 1997).
 P. 118.
 P. 203.
 Donna L. Feir and M. Christopher Auld, “Indian Residential Schools: Height and Body-mass post 1930,” Canadian Journal of Economics 54 (2021), 126-163. Indian residential schools: Height and body mass post‐1930 (wiley.com).
 Ioannidis, op. cit., corollary 5.
 Ioannidis, op. cit, Corollary 4.
 Ioannidis, op. cit., Corollary 1.
 Rodney A. Clifton and Mark DeWolf, “Putting the TRC Report into Context,” in Clifton and DeWolf, eds., From Truth Comes Reconciliation (Winnipeg: Frontier Centre for Public Policy, 2021), p. 37.
 Adding amounts mentioned in ibid plus $20 billion in Agreements-in-Principle reached on compensation and long-term reform of First Nations child and family services and Jordan's Principle - Canada.ca.